Search papers, labs, and topics across Lattice.
This paper introduces a variational latent equilibrium (VLE) framework to approximate backpropagation through time (BPTT) in a biologically plausible manner for learning complex spatiotemporal patterns. The approach leverages a prospective energy function of neuronal states to derive real-time error dynamics and, with modifications, achieves fully local (in space and time) credit assignment. The VLE framework unifies and extends previous approaches based on energy conservation and extremal action, offering a potential blueprint for physical circuits implementing spatiotemporal deep learning.
Bridging the gap between deep learning and neuroscience, this work presents a biologically plausible alternative to backpropagation through time, potentially unlocking new avenues for brain-inspired AI.
Brains remain unrivaled in their ability to recognize and generate complex spatiotemporal patterns. While AI is able to reproduce some of these capabilities, deep learning algorithms remain largely at odds with our current understanding of brain circuitry and dynamics. This is prominently the case for backpropagation through time (BPTT), the go-to algorithm for learning complex temporal dependencies. In this work we propose a general formalism to approximate BPTT in a controlled, biologically plausible manner. Our approach builds on, unifies and extends several previous approaches to local, time-continuous, phase-free spatiotemporal credit assignment based on principles of energy conservation and extremal action. Our starting point is a prospective energy function of neuronal states, from which we calculate real-time error dynamics for time-continuous neuronal networks. In the general case, this provides a simple and straightforward derivation of the adjoint method result for neuronal networks, the time-continuous equivalent to BPTT. With a few modifications, we can turn this into a fully local (in space and time) set of equations for neuron and synapse dynamics. Our theory provides a rigorous framework for spatiotemporal deep learning in the brain, while simultaneously suggesting a blueprint for physical circuits capable of carrying out these computations. These results reframe and extend the recently proposed Generalized Latent Equilibrium (GLE) model.
